Apply sequential learning to improve the discovery, development, and approval of treatments for unmet medical needs. Projects include designing adaptive clinical trials to help accelerate treatment development for rare diseases and efficiently screening large chemical libraries to identify promising molecules.
Develop learning theory and algorithms to identify the best choice among many alternatives while minimizing experimental or computational effort. Tackle fundamental trade-offs and challenging environments, including large-scale problems, simulation-input uncertainty, and heavy-tailed randomness.
Use operations research (OR) to make AI experimentation and evaluation more efficient, and explore how AI can support decision-making. Projects include learning algorithms that reduce token usage and computational cost in large-scale AI experiments, AI-generated personas as simulated participants in selection and optimization studies, and new approaches to evaluating large language models.
Develop digital representations of complex systems that evolve with data and support experimentation, prediction, and decision-making. Explore how simulation, learning, and optimization can work together to understand system behavior, test interventions, and improve decisions under uncertainty.
Study when to experiment, what to learn, and when to act. Explore how the value of information, the cost of experimentation, and the consequences of delay shape learning strategies and the allocation of resources across competing opportunities.
Explore how learning, simulation, and optimization can address the needs of people with rare diseases, ageing populations, and communities facing gaps in access to care. Questions include how to guide healthcare innovation, allocate limited resources, and design more responsive, equitable, and sustainable care systems.
Explore how AI can expand the ways we learn, model, and make decisions, and how learning theory and operations research can improve AI systems. Questions include efficient and reliable AI experimentation, meaningful evaluation, human–AI collaboration, and decision-making with evolving AI capabilities.
03 / Acknowledgements
We thank the coauthors and colleagues whose ideas and contributions shape this work. Paper-specific funding and support are acknowledged in the individual publications.
Meet our collaborators